Newsletter #98

Release of HDF Java Products Version 2.4

We are pleased to announce the release of the HDF Java Products, version 2.4, built with HDF 4.2r2 and HDF5 1.6.6.

This version can be obtained from the HDF Java Products web page at:

   https://support.hdfgroup.org/products/java/hdf-java-html/

The ftp server location of the HDF Java Products source code and pre-built binaries is:

   ftp://ftp.hdfgroup.org/HDF5/hdf-java/

New Features

There are many new features in this release. Following is a list of the major new features in HDFView:

• Support for compound datatypes containing 2D arrays (or greater)
• Ability to create/display named datatypes in an HDF5 file
• Ability to copy an object to the same group with a different name
• Ability to create a dataset with an enum type/named datatype
• Support for "filtering fillvalue" in the Image Viewer
• Ability to display the actual palette that results from palette manipulation
• Ability to create a chunked dataset with a compound datatype
• Ability to handle a large number of 8k x 8k 3D images by reuse of the memory data buffer
• Addition of the -geometry switch to set the window size and location
• Support for large fonts in GUI components
• Ability to grab and drag to move/browse an image
• Addition of the autogain algorithm for image Brightness/Contrast

Platforms Supported

Version 2.4 of the HDF Java Products is supported on the following platforms:

32-bit Java 2 SDK:

• Linux
• Solaris
• Mac PowerPC
• Mac Intel
• Windows (Vista/XP/2000)

64-bit Java 2 SDK:

• Linux 64-bit AMD
• Solaris 64-bit

Bug Fixes

There are many bug fixes in HDF Java 2.4. One of the major fixes was the following memory leak:

In HDF Java 2.3 (and earlier) an attribute and datatype were left open when the file structure was retrieved from a file. The following code could build up the memory leak until the machine was out of memory.

while (true)
{
    H5File f=new H5File(fname, H5File.READ);
    f.open();
    f.close();
}

Another memory leak that was fixed was with compound datasets. When a compound dataset is opened, the datatypes of the compound fields are stored in memory so that they can be reused for better performance. However, these datatypes were not closed when the file was closed. The following code could make the JVM run out of memory.

while (true)
{
    final H5File file = new H5File(fname, H5File.READ);
    final Dataset dset = (Dataset)file.get("/Table0");
    dset.init();

    file.close();
}

Other Enhancements

• Test Suite:
  Using the junit test frame, we build a test suite to test all the public APIs in the object package (ncsa.hdf.object) and the HDF5 object package (ncsa.hdf.object.h5). The test suite was added into the source configuration to make it easy to run with "make check".
• Improved Documentation:
  All of the public APIs in the object package (ncsa.hdf.object) and the HDF5 object package (ncsa.hdf.object.h5) are fully documented.

Please let The HDF Group Helpdesk know if you have any questions.